Re: Setting the state of the trout pond vs invoking the CleanTroutPon d method (was RE: Dead trout)

On Wed, Feb 19, 2003 at 01:35:59PM -0500, Christopher B Ferris wrote:
> I can't help myself... someone stop me before I shoot again...

Taken to www-archive ...

> Mark... please remind me how an autonomous (e.g. no human intervention) 
> agent gains 
> understanding and capacity to turn off your lightbulb or empty your trout 
> pond of floaters without
> being a specialized client with all sorts of built-in knowledge?

It's been hardcoded.

> How is it any different to have intimate knowledge and understanding about 
> how/what manipulation of 
> state can result in a desired outcome is any different than having 
> understanding and built-in knowledge 
> of what messages to send to some endpoint? 

I don't know what you mean by that last part.  As I see it, there are
two different things that both types of clients have to know;

- how the returned data reflects the state
- what API to use to access and manipulate that state

The first is identical in both the REST and WSA cases.  The second is
different, as the REST based solution uses an additional architectural
constraint which induces additional properties not exhibited (to the
same extent) by WSA.

I've gone on before about the value of visibility for Internet scale
systems, and won't do so again here.  But you really have to see the
value in it in order to understand how I go from "different solutions
have different properties" to "Web services won't work on the
Internet".  But if you haven't made that leap, then all you can know
is that there *is* a difference between the two approaches, you just
can't predict the impact that difference will have, as I do when I
suggest that Web services will fail to be deployed in large numbers
on the Internet.

MB
-- 
Mark Baker.   Ottawa, Ontario, CANADA.        http://www.markbaker.ca
Web architecture consulting, technical reports, evaluation & analysis

Received on Wednesday, 19 February 2003 14:38:43 UTC